摘要
当数据缺失机制为非随机缺失(NMAR)时,线性回归模型中的参数估计是一个复杂的问题.本文采用贝叶斯(Bayesian)方法,并利用MCMC方法、选择模型和Gibbs抽样方法,求得了参数的Bayesian估计,用一个模拟例子求得参数的估计均值和方差说明了此种方法的可行性。
The parameter estimation of the linear regression model is a very complex problem when the data missing mechanism is NMAR. By using the methods of Bayesian, MCMC, selection model and Gibbs sampling, this paper get the Bayesian estimation of parameter. By using a simulated example to get the mean and variance of parameter estimation, this paper explains the feasibility of this kind of method.
